- Open Access
Energy density and weight change in a long-term weight-loss trial
© Flood et al; licensee BioMed Central Ltd. 2009
Received: 4 May 2009
Accepted: 14 August 2009
Published: 14 August 2009
Health risks linked to obesity and the difficulty most have in achieving weight loss underscore the importance of identifying dietary factors that contribute to successful weight loss.
This study examined the association between change in dietary energy density and weight loss over time. Subjects were 213 men and women with BMI of 30–39 kg/m2 and without chronic illness enrolled in 2004 in a randomized trial evaluating behavioral treatments for long-term weight loss. Subjects completed a 62-item food frequency questionnaire at baseline and at 6, 12, and 18 months.
Pearson correlations between BMI and energy density (kcals/g of solid food) at baseline were not significantly different from zero (r = -0.02, p = 0.84). In a longitudinal analysis, change in energy density was strongly related to change in BMI. The estimated β for change in BMI (kg/m2) of those in the quartile representing greatest decrease in energy density at 18 months compared to those in the quartile with the least was -1.95 (p = 0.006). The association was especially strong in the first six months (estimated β = -1.43), the period with greatest weight loss (mean change in BMI = -2.50 kg/m2 from 0–6 months vs. 0.23 kg/m2 from 12–18 months) and the greatest contrast with respect to change in energy density.
Decreased energy density predicted weight loss in this 18 month weight loss study. These findings may have important implications for individual dietary advice and public health policies targeting weight control in the general population
National Health and Nutrition Examination Survey (NHANES) data indicate that in excess of 30% of Americans over the age 20 are obese (i.e., have a body mass index or BMI in excess of 30 kg/m2). Furthermore, prevalence of obesity has shown rapid increase throughout the 1990's and has continued to rise unabated since then [1–3]. Given the well-established links between obesity and numerous chronic diseases (e.g., heart disease, diabetes, hypertension, and many cancers) [4–6], identifying successful strategies for achieving weight loss would have significant public health impact.
Dietary energy density, or an individual's average daily energy intake divided by the grams of food consumed per day, is a factor that could help explain the broader trends in obesity prevalence and that could also serve as a potentially important point of focus in efforts to achieve weight loss. At the population level, dietary energy density is associated with energy intake and weight status in a limited number of cross-sectional studies [7–10]. At the individual level, higher energy density diets in feeding studies are associated with greater energy intake but not with a different volume of food consumed. If changing to a diet comprised of low energy density foods results in fewer calories consumed, as these results suggest it should, then doing so should also result in a reduction in weight. Only a small number of longer-term studies have examined the question of changing energy density and its implications for weight loss [12–15]. In the only observational analyses from a study where energy density per se was investigated as a predictor of weight change, subjects with the greatest decrease in energy density achieved the greatest weight loss.
We conducted a longitudinal analysis of data collected from a randomized weight loss trial. It was our primary aim in these analyses to determine if a change in dietary energy density would result in weight loss over 18 months of follow-up, a follow-up period that extended beyond what has been reported in the small number of prior studies of energy density and weight loss. We had the secondary aim of analyzing the data at intermediate stages of follow up in order to explore the dynamics of the relationship between changes in energy density and weight loss and to provide insight into potential mechanistic explanations for the associations we observed.
Subjects: Participants in the study were 100 men and 113 women recruited in 2004 and 2005 to participate in the Lose It Forever (LIFE) Study, a randomized clinical trial designed to compare the efficacy of prescribing different behavior combinations over time during therapy as a means of achieving and then maintaining weight loss.
Participants were recruited by general mass media advertising and specialized advertising targeting men and minorities. To be eligible for the study, subjects had to be at least 18 years of age, have a body mass index (BMI) between 30 and 39 kg/m2, be free from serious medical conditions or other consideration that would contraindicate treatment (i.e., participation in a weight loss intervention), agree to be randomized to either of two treatment groups, and provide their informed consent to participate. The study protocol was approved by the Institutional Review Boards at both the University of Minnesota and University of Washington.
Since the primary endpoint for this analysis was weight loss at 18 months of follow-up, we included only those subjects who had weight measurements at the 18 month clinic visit (see below) in the final analyses as well as in the cross-sectional analyses at baseline and in the analyses of intermediate time points. Doing so resulted in an analytic sample of 155 subjects (73% of those randomized). Including all subjects for the baseline cross-sectional analyses or including subjects with weight measures at the 6 or 12 month clinic visits but not the 18 month visit did not produce results materially different from those using only the subjects who had weight measures at 18 month completion point (data not shown).
LIFE Study Intervention: The present analysis uses the LIFE Study data in an observational manner, though the larger study used an experimental design as described here. Following enrollment, participants were randomized to one of two treatment groups. The treatment in both groups was conducted in small groups of 11 to 21 individuals at the Epidemiology Clinical Research Center at the University of Minnesota. The format of both treatment groups was comprised of a presentation by treatment staff of recommendations for changes in diet and physical activity behaviors necessary for successfully controlling weight, interactive discussion of these behavioral goals and strategies to achieve them, and homework assignments to be completed between sessions to reinforce learning.
Content for the Standard Behavioral Treatment (SBT) group was closely modeled after prior work of the investigatorsand closely resembled the lifestyle intervention programs administered in recent successful clinical trials[18, 19]. It was comprised of 26 weekly meetings over the first six months, bi-weekly meetings between months 6 and 12, and monthly meetings between months 12 and 18. Participants were given a calorie intake goal of 1,200, 1,500 or 2,000 kilocalories per day, depending on initial body weight, and were progressively advised to increase their level of moderate intensity physical activity (the primary exercise recommended was walking) to a total of one hour per day. The therapy in the Maintenance-Tailored Treatment (MTT) incorporated the same general diet and exercise goals but emphasized greater variety in both format and content. Energy density per se was not the focus of any of the weight loss units for either treatment arm. Encouragement to eat lower energy foods was a recurring theme, however, and individuals who act on this encouragement will often eat a lower energy density diet.
Outcome and Covariate Measures: The primary outcome was change in BMI. Study personnel used standardized procedures to measure subjects: Weight measurements were repeated, in light clothing without shoes, on an electronic scale during assessment sessions (in the morning, fasted) at baseline and at 6, 12, and 18 months of follow-up. Height was measured at the baseline assessment visit using a wall mounted stadiometer. LIFE Study personnel called participants prior to each six-month outcome assessment session to encourage attendance. In the event of a no-show at the assessment session, study personnel made multiple reschedule attempts.
Demographic characteristics of age, gender, and ethnicity were assessed at baseline. The Paffenbarger Activity Questionnaire, a well known assessment instrument with established validity and reliability [20–22], was used to estimate usual physical activity.
Dietary Assessment: Subjects completed the 62-item Block Food Frequency Questionnaire (FFQ) to assess usual intake in the previous 6 months. Subjects completed the questionnaires individually and in a self-administered manner (i.e., without coaching from study personnel, although study personnel did review the FFQs for completeness accuracy when participants submitted them). Detailed descriptions of this FFQ and its validity have appeared elsewhere [23–25]. Software designed for this FFQ yielded estimates of daily intakes for total energy, macronutrients, and micronutrients, and grams of total foods and individual food items.
Baseline (unless otherwise indicated) characteristics of LIFE Study subjects according to quartile of change in energy density from baseline to 18 months of follow-up (all values are percents or means in units listed)
Quartile of Change in Energy Density from Baseline to 18 Months (N = 155)
Gender (% male)
Ethnicity (% white)
Treatment group (% randomized to MTT)
Smokers (current/former) (%)
More than HS education (%)
Energy density at baseline (kcals/g)
Energy density at 18 months (kcals/g)
Energy at baseline (kcals/day)
Quantity of food at baseline (g/day)
Physical activity at baseline (kcals/week)
Physical activity at 18 months (kcals/week)
Percent of Energy from Fat
Baseline characteristics for the LIFE study subjects, by quartile of change in energy density from baseline to 18 months, appear in Table 1. Interestingly, those with the greatest decrease in energy density from 0 to 18 months had the highest energy density at baseline, and across quartiles of change in energy density, the baseline energy density went down monotonically. By contrast, the quantity of food consumed was essentially the same across quartiles, and thus the energy consumption at baseline was lowest in the top quartile, where the energy density was also lowest. An equally interesting observation was that at 18 months, quartile 1, the quartile with the largest decrease in energy density, had the lowest mean energy density and across quartiles the energy density values increased, exactly the opposite of what we observed for baseline when quartile 1 had the highest mean energy density.
Cross-sectional correlations among baseline measures of energy density, total energy intake, and BMI (N = 155)
Pearson Correlation Coefficients
Energy Density (kcals/g)
All Food – Excluding Non-dairy Beverages
p = 0.35
p = 0.89
p = 0.84
Total Energy (kcals)
p = 0.87
p = 0.27
p < 0.0001
Changes in energy densitya as predictors of changes in BMIb and energy density
Estimated β for Δ BMI (kg/m 2 ) resulting from Δ Energy Density (Q1 vs. Q4) b
p < 0.0001
p = 0.01
p = 0.15
p = 0.006
Mean Δ BMI
(N = 153)
(N = 149)
(N = 151)
(N = 155)
Mean Δ ED
Q1 (N = 39)
Q4 (N = 38)
That the relationship between change in energy density and weight change would decline over time may at first appear puzzling, but it is important to note that the comparison that formed the basis of the β estimates (Q1 vs. Q4) during the different phases of follow-up represented much different levels of change in both the energy density and BMI variables during the different time periods. During the first six months of follow-up, Q1 represented a mean decrease in energy density of 1.39 kcal/g while Q4 represented a mean decrease of only 0.05 kcal/g. By contrast, in the 12–18 month period, Q1 represented a decrease in mean energy density of just 0.05 kcal/g while Q4 represented an increase of 0.38 kcal/g. Thus the difference between Q1 and Q4 was more than 3 times greater during the 0–6 month period than it was in the 12–18 month period. The absolute decrease in energy density in Q1 was more than 20 times greater during the early period of follow-up compared to the 12–18 month period.
In an effort to explore mechanistic explanations for the association between decreasing energy density and weight loss, we conducted a secondary analysis using a model that further controlled for total energy consumed. Contrary to expectations, doing so had no appreciable affect on the β estimates when comparing subjects with the largest reported decreases in energy density (Q1) to those who reported no change or a slight increase in energy density (Q4) (data not shown).
Mean daily volume of food and energy consumed by quartile of baseline to 18 months change in energy densitya
Grams of Food Consumed per Day (Std Dev)b
0–18 Month Δ Energy Density
Q1 – decrease (N = 39)
Q2 (N = 38)
Q3 (N = 40)
Q4 – increase (N = 38)
Kcals of Energy per Day (Std Dev) c
0–18 Month Δ Energy Density
Q1 – decrease (N = 39)
Q2 (N = 38)
Q3 (N = 40)
Q4 – increase (N = 38)
In this 18 month weight loss intervention study, we observed that subjects with the greatest degree of change in energy density did not change the amount (i.e., weight) of food they consumed but did have notable changes in energy intake. These observations are consistent with what Rolls and colleagues have observed in short-term feeding studies where study subjects consuming diets differing in energy density ate a constant volume of food but the low energy-density diet subjects consumed significantly fewer calories[10, 11, 27, 28]. After controlling for changes in physical activity during the intervention and for other potential confounders, we further found that the changes we observed in energy density (and in mean daily energy consumption) were associated with significant changes in weight.
Some of our findings were somewhat unexpected, however. For example, despite prior studies consistently showing an association between energy density of the diet and BMI [7–10], we found there was no correlation at baseline between energy density and BMI among LIFE Study subjects. A key difference in our study may be the range of values for BMI among study subjects at enrollment. By design, everyone in the LIFE Study was obese (i.e., had a BMI in excess of 30 kg/m2), but no one had a BMI above 39 kg/m2. By restricting the range of values for BMI in this way, we may have obscured the cross-sectional association most previous investigators have observed.
A second unexpected finding was that the association between change in energy density and change in weight or BMI was different at different phases of follow-up. During periods of active weight loss, that weight loss was strongly correlated with decreased energy density, but later in follow-up, the association was markedly attenuated. Our evaluation of the magnitude of change in energy density during each of these time periods showed that large changes in energy density, both in absolute magnitude (i.e., -1.39 kcal/g for those in the quartile with the greatest decline in energy density from 0–6 months compared to -0.05 kcal/g for those showing the greatest decline in energy density from 12–18 months) and in relative difference (1.34 kcal/g difference in mean change in energy density for Q1 vs. Q4 from 0–6 months compared to 0.43 kcal/g from 12–18 months) suggest that this apparently changing association may simply be a function of the greater range in exposure levels early in the follow-up period compared to later.
The underlying theory as to why decreasing energy density would result in weight loss or lower BMI is that food volume is an important determinant of food intake. Decreased energy density in the context of constant volume of food consumption would necessarily involve less energy intake, just as we observed in the LIFE Study. If we control for changes in physical activity, then the lower energy intake should result in weight loss meaning the energy density effect is mediated by total energy consumption. And yet, in secondary analyses, when we adjusted for total energy, the estimated β's were largely unaffected (data not shown). If energy were truly the mechanism by which changes in energy density impacted BMI, then controlling for energy should have eliminated the effect of changing energy density. That it did not could potentially be explained as evidence that the association we saw was the result of confounding by some unmeasured or imperfectly measured factor.
Rolls and colleagues found results consistent with ours in a 12 month study of 200 overweight and obese individuals. In this study all subjects were on a calorie restricted diet, but after 1 year, subjects consuming a low energy density soup twice a day as compared to subjects eating a high energy density snack with the same energy content as the soup had significantly greater weight loss.
That the association of energy density with weight loss and energy intake was strongest if we defined it only in terms of solid food consumed (i.e., if we excluded liquid food), is worth noting. This result suggest that the variation in energy consumption that comes from changes in energy density is primarily a function of solid foods, which should not be surprising as liquid foods as a rule are low in energy density. Furthermore, Rolls and colleagues have found that varying the amount of drinking water with meals did not affect volume of food consumption suggesting the effects of changes in energy density would be concentrated at the level of solid foods. As such, it is clear that substituting low energy-density solid foods in place of high energy density solid foods will impact energy intake, but the effects of liquid foods are much less obvious. Our results are consistent with these previous observations.
An interesting secondary observation in this study was the poor weight loss outcome in subjects starting with low energy density diets at baseline. There are several possible explanations for this result. The first is simply that subjects eating high energy density diets had substantial room for improvement on that front, and when they did reduce the energy density of their diets, their energy intake dropped in turn. For subjects with low energy density at baseline, however, no additional improvement of this type was possible. For the latter group of subjects, reducing energy intake could not happen through changes in energy density of the diet, and alternative strategies (i.e., portion control) would be required for effective weight loss. Alternatively, the low energy density diets reported in the quartile showing the least change in energy density over the course of the study might be erroneous and as such reflect poor dietary self awareness. Given that all subjects had starting BMIs of between 30 and 39 kg/m2 and thus had similar degrees of energy imbalance, it is perhaps not unreasonable to consider that those reporting the lowest energy density diets were in fact underestimating the amount of high energy density foods consumed relative to the rest of the study population. Thus reporting a low energy density diet at baseline may be an indicator of poor behavioral compliance (and hence little improvement in energy balance) among subjects who do not recognize the energy density of their diets at the outset. These explanations are strictly hypothetical, however, and require further investigation in future studies before any conclusions are possible.
The use of a food frequency questionnaire to assess energy density of diet might be a point of criticism for this study as most investigations of energy density have been short term feeding studies with tightly regulated (and measured) food intakes. There is no question that food frequency questionnaires measure diet with error, so much so that some have suggested dispensing with them altogether. But this criticism is aimed primarily at the failure of studies using FFQs, due to the substantial random error inherent in those instruments, to find positive associations. In the case of the present analysis, we found that greater reduction in energy density resulted in greater weight loss, and dismissing that result on the grounds that the FFQ is inadequate to assess diet properly requires a belief that there is not simply random error but rather reporting bias as a function of success in weight loss. In other words, there would need to have been systematic over reporting of "good" habits (eating low energy density foods) in those who were successful at weight loss. While we cannot exclude this possibility, the reverse scenario (social desirability-motivated over-reporting of "good" behaviors in those who failed to lose weight) seems much more likely. That we found a positive result despite the limitations of the FFQ is notable.
A final limitation of the study may be in the demographic makeup of the study population. Results from this mostly white, generally well-educated population (in excess of 97% had more than a high school education) may not be generalizable to the broader public. Certainly, it is reasonable to consider if achieving these changes in dietary energy density would be similarly practical in different populations, but it is not immediately clear how change in energy density per se could have different effects in populations with different ethnic or educational profiles. Nonetheless, some caution is warranted in interpreting these results.
In summary, decreases in energy density predicted weight loss in an 18 month observational analysis. Changes in energy density predicted weight loss most strongly in the first 6 months of follow-up. Diminished predictive value of energy density later in follow-up may be due to smaller changes in energy density during this period making for a less sensitive analysis. Volume of food consumed did not change with decreasing energy density suggesting energy intake went down, but curiously, controlling for energy intake did not eliminate the effect of changing energy density on weight loss.
The research described in this paper was supported by NIH K07 CA108910-01A1 and NIH R01 DK064596. The funding institutions had no role in study design; in the collection, analysis, and interpretation of data; in the writing of the manuscript; and in the decision to submit the manuscript for publication.
- Flegal KM, Carroll MD, Ogden CL, Johnson CL: Prevalence and trends in obesity among US adults, 1999–2000. JAMA. 2002, 288: 1723-7. 10.1001/jama.288.14.1723.View ArticleGoogle Scholar
- Mokdad AH, Ford ES, Bowman BA, Dietz WH, Vinicor F, Bales VS, Marks JS: Prevalence of obesity, diabetes, and obesity-related health risk factors, 2001. JAMA. 2003, 289: 76-79. 10.1001/jama.289.1.76.View ArticleGoogle Scholar
- Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal KM: Prevalence of overweight and obesity in the United States, 1999–2004. JAMA. 2006, 295: 1549-55. 10.1001/jama.295.13.1549.View ArticleGoogle Scholar
- World Cancer Research Fund, American Institute for Cancer Research: Food, Nutrition, Physical Activity, and the Prevention of Cancer: a Global Perspective. 2007, Washington, DC: AICRGoogle Scholar
- WHO: Obesity: preventing and managing the Global epidemic – Report of a WHO consultation on obesity, 3–5 June 1997, Geneva. 2003, Geneva, Switzerland: World Health OrganizationGoogle Scholar
- Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ: Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med. 2003, 348: 1625-38. 10.1056/NEJMoa021423.View ArticleGoogle Scholar
- Murakami K, Sasaki S, Takahashi Y, Uenishi K: Dietary energy density is associated with body mass index and waist circumference, but not with other metabolic risk factors, in free-living young Japanese women. Nutrition. 2007, 23: 798-806. 10.1016/j.nut.2007.08.014.View ArticleGoogle Scholar
- Kant AK, Graubard BI: Energy density of diets reported by American adults: association with food group intake, nutrient intake, and body weight. Int J Obes (Lond). 2005, 29: 950-6. 10.1038/sj.ijo.0802980.View ArticleGoogle Scholar
- Stookey JD: Energy density, energy intake and weight status in a large free-living sample of Chinese adults: exploring the underlying roles of fat, protein, carbohydrate, fiber and water intakes. Eur J Clin Nutr. 2001, 55: 349-59. 10.1038/sj.ejcn.1601163.View ArticleGoogle Scholar
- Ledikwe JH, Blanck HM, Kettel Khan L, Serdula MK, Seymour JD, Tohill BC, Rolls BJ: Dietary energy density is associated with energy intake and weight status in US adults. Am J Clin Nutr. 2006, 83: 1362-8.Google Scholar
- Bell EA, Rolls BJ: Energy density of foods affects energy intake across multiple levels of fat content in lean and obese women. Am J Clin Nutr. 2001, 73: 1010-8.Google Scholar
- Greene LF, Malpede CZ, Henson CS, Hubbert KA, Heimburger DC, Ard JD: Weight maintenance 2 years after participation in a weight loss program promoting low-energy density foods. Obesity (Silver Spring). 2006, 14: 1795-801. 10.1038/oby.2006.207.View ArticleGoogle Scholar
- Rolls BJ, Roe LS, Beach AM, Kris-Etherton PM: Provision of foods differing in energy density affects long-term weight loss. Obes Res. 2005, 13: 1052-60. 10.1038/oby.2005.123.View ArticleGoogle Scholar
- Ello-Martin JA, Roe LS, Ledikwe JH, Beach AM, Rolls BJ: Dietary energy density in the treatment of obesity: a year-long trial comparing 2 weight-loss diets. Am J Clin Nutr. 2007, 85: 1465-77.Google Scholar
- Blackburn GL, Kanders BS, Lavin PT, Keller SD, Whatley J: The effect of aspartame as part of a multidisciplinary weight-control program on short- and long-term control of body weight. Am J Clin Nutr. 1997, 65: 409-18.Google Scholar
- Ledikwe JH, Rolls BJ, Smiciklas-Wright H, Mitchell DC, Ard JD, Champagne C, Karanja N, Lin PH, Stevens VJ, Appel LJ: Reductions in dietary energy density are associated with weight loss in overweight and obese participants in the PREMIER trial. Am J Clin Nutr. 2007, 85: 1212-21.Google Scholar
- Jeffery RW, Wing RR, Sherwood NE, Tate DF: Physical activity and weight loss: does prescribing higher physical activity goals improve outcome?. Am J Clin Nutr. 2003, 78: 684-9.Google Scholar
- Rubin RR, Fujimoto WY, Marrero DG, Brenneman T, Charleston JB, Edelstein SL, Fisher EB, Jordan R, Knowler WC, Lichterman LC, Prince M, Rowe PM, DPP Research Group: The Diabetes Prevention Program: recruitment methods and results. Control Clin Trials. 2002, 23: 157-71. 10.1016/S0197-2456(01)00184-2.View ArticleGoogle Scholar
- Pi-Sunyer X, Blackburn G, Brancati FL, et al: Reduction in weight and cardiovascular disease risk factors in individuals with type 2 diabetes: one-year results of the look AHEAD trial. Diabetes Care. 2007, 30: 1374-83. 10.2337/dc07-0048.View ArticleGoogle Scholar
- Harris JK, French SA, Jeffery RW, McGovern PG, Wing RR: Dietary and physical activity correlates of long-term weight loss. Obes Res. 1994, 2: 307-13.View ArticleGoogle Scholar
- Paffenbarger RS, Wing AL, Hyde RT: Physical activity as an index of heart attack risk in college alumni. Am J Epidemiol. 1978, 108: 161-75.Google Scholar
- Sherwood NE, Jeffery RW, French SA, Hannan PJ, Murray DM: Predictors of weight gain in the Pound of Prevention study. Int J Obes Relat Metab Disord. 2000, 24: 395-403. 10.1038/sj.ijo.0801169.View ArticleGoogle Scholar
- Block G, Hartman AM, Dresser CM, Carroll MD, Gannon J, Gardner L: A data-based approach to diet questionnaire design and testing. American Journal of Epidemiology. 1986, 124: 453-469.Google Scholar
- Block G, Hartman AM, Naughton D: A reduced dietary questionnaire: Development and validation. Epidemiology. 1990, 1: 58-64. 10.1097/00001648-199001000-00013.View ArticleGoogle Scholar
- National Cancer Institute , Block Dietary Data Systems., Information Management Services Inc.: DIETSYS version 3.0 user's guide. 1994, Bethesda, MD: National Cancer InstituteGoogle Scholar
- Ledikwe JH, Blanck HM, Khan LK, Serdula MK, Seymour JD, Tohill BC, Rolls BJ: Dietary energy density determined by eight calculation methods in a nationally representative United States population. J Nutr. 2005, 135: 273-8.Google Scholar
- Bell EA, Castellanos VH, Pelkman CL, Thorwart ML, Rolls BJ: Energy density of foods affects energy intake in normal-weight women. Am J Clin Nutr. 1998, 67: 412-20.Google Scholar
- Rolls B, Bell E, Castellanos V, Chow M, Pelkman C, Thorwardt M: Energy Density but not Fat Content of Foods Affected Energy Intake in Lean and Obese Women. American Journal of Clinical Nutrition. 1999, 69: 863-871.Google Scholar
- Drewnowski A, Almiron-Roig E, Marmonier C, Lluch A: Dietary energy density and body weight: is there a relationship?. Nutr Rev. 2004, 62: 403-13. 10.1111/j.1753-4887.2004.tb00012.x.View ArticleGoogle Scholar
- Rolls BJ, Bell EA, Thorwart ML: Water incorporated into a food but not served with a food decreases energy intake in lean women. Am J Clin Nutr. 1999, 70: 448-55.Google Scholar
- Kristal AR, Peters U, Potter JD, Editorial: Is it time to abandon the food frequency questionnaire?. Cancer Epidemiol Biomarkers Prev. 2005, 14: 2826-8. 10.1158/1055-9965.EPI-12-ED1.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.